Re: [PATCH] Record and output access specifiers for nested typedefs

[Pedro Alves <palves@redhat.com>]

On 10/16/2017 10:32 PM, Keith Seitz wrote:
> @@ -13274,6 +13274,28 @@ dwarf2_add_typedef (struct field_info *fip, struct die_info *die,
>  
>    fp->type = read_type_die (die, cu);
>  
> +  /* Save accessibility.  */
> +  enum dwarf_access_attribute accessibility;
> +  struct attribute *attr = dwarf2_attr (die, DW_AT_accessibility, cu);
> +  if (attr != NULL)
> +    accessibility = (enum dwarf_access_attribute) DW_UNSND (attr);
> +  else
> +    accessibility = dwarf2_default_access_attribute (die, cu);
> +  switch (accessibility)
> +    {
> +    case DW_ACCESS_public:
> +      /* The assumed value if neither private nor protected.  */
> +      break;
> +    case DW_ACCESS_private:
> +      fp->is_private = 1;
> +      break;
> +    case DW_ACCESS_protected:
> +      fp->is_protected = 1;
> +      break;
> +    default:
> +      gdb_assert_not_reached ("unexpected accessibility attribute");
> +    }
> +

Please don't gdb_assert on invalid input.  Bogus input caused by
a buggy compiler shouldn't bring down gdb.

(I can't claim that I've really read the patch; I've barely
skimmed it, there may be other instances.  Please double check.)

Thanks,
Pedro Alves